top of page

Attribution and Dynamics in Online Advertising

Table of Content

  1. General Questions on Attribution and Dynamics in Online Advertising

    1. Do display ads influence search ad performance and conversions?

    2. Do online ads exhibit dynamic effects that improve effectiveness over time?

    3. How should attribution and dynamics impact online advertising metrics and budget allocation?


General Questions on Attribution and Dynamics in Online Advertising


Online advertising has become a pivotal component of marketing strategies across industries. As businesses increasingly allocate substantial budgets towards digital channels, a pressing need arises to comprehend the intricate dynamics and attribution effects that shape the success of these campaigns.

This post delves into the fundamental questions that underscore the importance of understanding attribution and dynamics in online advertising:


A. Do Display Ads Influence Search Ad Performance and Conversions?


One of the core inquiries revolves around the interplay between display advertising and search advertising. Specifically, marketers seek to unravel the extent to which display ads influence the performance of search campaigns, including:


  • Search ad clicks: Do display ad exposures drive consumers to click on search ads, thereby increasing click-through rates (CTRs) and potential conversions?

  • Search ad conversions: Beyond clicks, do display ads contribute to actual conversions from search campaigns, such as purchases, lead generations, or other desired actions?

  • Search funnel progression: How do display ads impact the various stages of the consumer journey, from initial awareness to active search and eventual conversion?


Answering these questions is crucial for optimizing cross-channel marketing strategies and accurately attributing credit to each touchpoint along the customer's path to conversion.


B. Do Online Ads Exhibit Dynamic Effects that Improve Effectiveness Over Time?


Another critical area of investigation focuses on the dynamic effects of online advertising. Specifically:


  • Carryover effects: Do the impacts of online ads persist and accumulate over time, even after the initial ad exposure? If so, how long do these carryover effects last?

  • Wear-in effects: Is there a lag between ad exposure and its impact on consumer behavior, where the effects "wear in" gradually over time?

  • Long-term effectiveness: How do the long-term effects of online ads compare to their immediate or short-term impacts? Are there significant differences in effectiveness over various time horizons?


Understanding these dynamic patterns is essential for accurately measuring the true return on investment (ROI) of online advertising campaigns and optimizing ad timing and frequency.


C. How Should Attribution and Dynamics Impact Online Advertising Metrics and Budget Allocation?


The insights gained from studying attribution and dynamics have profound implications for how marketers evaluate and optimize their online advertising efforts, including:

Metric

Standard Approach

Attribution & Dynamics Approach

Cost per Acquisition (CPA)

Based solely on last-click or last-touch attribution

Accounts for cross-channel impacts and long-term effects

Return on Ad Spend (ROAS)

Calculated using immediate, short-term conversions

Incorporates delayed conversions and carryover effects

Conversion Rate

Attributes conversion to the final ad touchpoint

Distributes credit across multiple touchpoints

Budget Allocation

Relies on static, snapshot metrics

Optimized based on long-term, cross-channel effectiveness

Failure to account for attribution and dynamics can lead to:


  • Inaccurate performance measurement and ROI calculations

  • Suboptimal budget allocation across channels and campaigns

  • Missed opportunities to capitalize on synergies and compounding effects

  • Premature termination of effective campaigns due to underestimated long-term impact


As businesses strive to maximize the impact of their online advertising investments, addressing these fundamental questions becomes paramount. By gaining a comprehensive understanding of attribution and dynamics, marketers can unlock valuable insights, refine their strategies, and ultimately drive superior returns on their digital marketing efforts.


Harvard Study on Search and Display Ad Interaction


The Harvard study by Kireyev, Pauwels, and Gupta (2013) provides valuable insights into the interaction between paid search and display advertising in driving consumer conversions. The researchers developed a multivariate time series model using data from a large commercial bank on its online marketing spend, ad impressions/clicks, and customer acquisitions over a 1-year period.


Methodology


The study employed a comprehensive persistence modeling approach to capture the complex dynamics and interdependencies in online advertising. Specifically:


  • Granger causality tests identified which variables should be treated as endogenous in the model.

  • Unit root tests determined which variables exhibited non-stationary behavior and should enter in differences.

  • Cointegration tests uncovered stationary linear combinations representing long-run equilibrium relationships.


Based on these tests, the researchers specified a vector error correction (VEC) model with all variables as endogenous, allowing for rich interactions.


Key Findings of the Harvard Study on Search and Display Ad Interaction


The VEC model enabled the authors to derive several important findings through impulse response analysis:


1. Display Ads Increase Search Conversions


A key finding was that display ad impressions significantly increased paid search conversions, but this effect was not immediate. It took around 2 weeks for display ads to start positively impacting search ad conversion rates.

This highlights the importance of accounting for cross-channel attribution effects rather than crediting only the last click.


2. Strong Dynamics and Carryover Effects


Both search and display ads exhibited strong positive carryover effects that improved their effectiveness and return on investment over time. Ignoring these dynamic effects can lead to undervaluing the long-term impact of online ads.


3. Display Increases Search Costs


In addition to increasing search conversions, greater display ad exposure also drove more search ad clicks and costs. So the overall impact of display needs to account for this added search spend.


4. Revised Performance Metrics


After accounting for attribution to search, carryover effects, and added search costs, the study found:


  • Each $1 spent on display ads yielded $1.24 in revenue

  • Each $1 on search ads yielded $1.75 in revenue


This contrasted sharply with estimates based on standard last-click attribution metrics used by the bank.


5. Optimal Budget Reallocation


The revised performance metrics had major implications for optimal budget allocation between search and display:


  • Despite display's attribution benefit, the strong dynamic effects for search called for increasing the search budget share by up to 36%.

  • The display budget share should be reduced by 31%.


This reallocation was driven by the much higher advertising elasticities for search after properly accounting for dynamics and cross-channel effects.


The authors conducted several robustness checks on their modeling approach and results:


  • Estimating a basic VAR model showed qualitatively similar short-term and wear-in effects, confirming the non-stationarity only impacted long-run behavior.

  • Variance decompositions confirmed display impressions significantly drove search impressions (40%), clicks (17%), and conversions (16%).

  • Using different shock identification schemes did not materially impact the impulse response estimates.


The rigorous econometric methodology, extensive robustness analysis, and use of granular marketing data from a major firm lends strong credibility to the study's findings. By quantifying the complex interplay between search and display ads, the study provides a framework for resolving attribution problems and optimizing budget allocation across digital channels based on their long-term effectiveness.


Nielsen Study on New Media's Brand Impact

The Nielsen study aimed to quantify the impact of emerging digital media formats like podcasts, influencer marketing, and branded content on key marketing metrics such as brand awareness, brand recall, and return on investment (ROI). The large-scale study analyzed over 1,000 campaigns across these new media channels spanning multiple industries and brands.

Groups were surveyed before and after the campaign to measure changes in key brand metrics like:


  • Unaided Brand Awareness: Percentage who mentioned the brand when asked an open-ended question about brands in the category.

  • Aided Brand Awareness: Percentage correctly identifying the brand when shown visual logos/assets.

  • Brand Favorability: Percentage who rated the brand favorably on a scale.

  • Purchase Intent: Percentage likely to purchase the brand's products.


Key Findings of the Nielsen Study on New Media's Brand Impact


The study yielded several important insights regarding the ability of new media formats to drive upper-funnel brand marketing objectives:


1. Significant Brand Awareness Impact


Across the 1,000+ campaigns analyzed, Nielsen found that the median brand awareness lift among consumers exposed to podcasts, influencers, or branded content was over 70% higher compared to the control group.

This highlights the powerful ability of these emerging channels to increase brand salience and top-of-mind recall, even for consumers not actively seeking out brand information.


2. Variance in Effectiveness


While the median brand awareness lift was 70%, Nielsen observed significant variance in effectiveness across campaigns. The top quartile of campaigns achieved awareness lifts over 150%, while the bottom quartile saw negligible or even negative lifts.

This variance was driven by factors like:

  • Content quality and creativity

  • Audience targeting and relevance

  • Integration with other media channels

  • Clarity of brand messaging and presence


3. Impact on Lower-Funnel Metrics


In addition to brand awareness, the study also analyzed the impact on metrics lower in the marketing funnel:

Metric

Median Lift

Brand Favorability

28%

Purchase Intent

18%

Short-Term ROI

2.7x

While not as pronounced as brand awareness, Nielsen found meaningful lifts in brand perception, purchase intent, and short-term return on ad spend from new media campaigns.

This suggests that these channels can effectively drive outcomes beyond just awareness when executed properly and complemented with mid/lower-funnel tactics.


4. Importance of Attention and Viewability


A key driver of effectiveness was the ability to capture consumer attention and ensure the brand messaging was viewable. Campaigns with higher viewability rates and longer average attention times tended to see larger lifts across all brand metrics analyzed.

This underscores the importance of using new media formats and distribution channels that foster active engagement rather than just passive exposure.


The Nielsen study provides robust evidence that digital platforms like podcasts, influencers, and branded content can be powerful vehicles for building brand awareness and driving marketing impact when leveraged strategically.

By using an advanced methodology to isolate the true effects of new media exposure, the study highlights both the significant potential of these channels as well as the key factors that differentiate high and low performing campaigns. As consumer attention becomes increasingly fragmented across digital media, these insights can help brands navigate the evolving landscape and optimize their media mix for maximum brand building impact.


Lumen/Teads/Dynata Study on Attention and Brand Effects


This study was a collaborative effort between Lumen Research, Teads, and Dynata, aimed at quantifying the relationship between attention metrics for digital advertising and the resulting brand effects. It involved a large-scale meta-analysis spanning campaigns across 14 major advertisers in 2022 and 2023.The analysis combined two key data sources:


  1. Third-Party Attention Data

  • Collected by Lumen across over 500 digital ad campaigns

  • Measured viewable impressions and attentive seconds per impression

  • Covered various ad formats like display, video, mobile, etc.

  1. Brand Lift Studies

  • Conducted by Dynata for the same set of campaigns

  • Surveyed a treatment group exposed to the ads and a control group

  • Measured impact on brand awareness, favorability, consideration, purchase intent


By combining these robust datasets, the researchers could analyze the correlation between objective attention metrics and the resulting brand KPIs at a very granular level.


Key Findings of the Lumen/Teads/Dynata Study on Attention and Brand Effects


The meta-analysis uncovered several important insights regarding the relationship between attention and brand effects:


1. Clear Correlation Between Attention and Brand Metrics


Across the 500+ campaigns studied, the analysis found a statistically significant and "clear correlation" between attentive ad exposure and positive shifts in upper-funnel brand metrics like:

  • Unaided Brand Awareness

  • Aided Brand Awareness

  • Brand Favorability

  • Brand Consideration


The higher the attentive seconds per impression, the greater the observed brand lift tended to be.


2. Threshold Effects for Lower-Funnel Metrics


While attention correlated with upper-funnel metrics across the board, the study found evidence of threshold effects for lower-funnel metrics like purchase intent:


  • Minimal lift was seen at low attention levels (<5 attentive seconds)

  • A steep increase in purchase intent lift occurred between 5-10 attentive seconds

  • Beyond 10 attentive seconds, further lift diminished


This suggests that driving lower-funnel KPIs requires a higher bar for attentive ad exposure compared to just building awareness.


3. Outperformance of Attentive Impressions


The analysis compared the brand impact of "attentive impressions" (defined as >5 attentive seconds) versus standard viewable impressions:

Metric

Lift from Viewable Imps

Lift from Attentive Imps

Unaided Awareness

9%

24%

Brand Favorability

6%

17%

Purchase Intent

3%

12%

Attentive impressions consistently drove 2-4x higher brand lift across the board compared to just viewable impressions.


4. Creative Optimization Opportunities


While overall attention levels were a key driver, the study also found that creative execution factors like:

  • Visual salience and clutter

  • Branding clarity and consistency

  • Storytelling and emotional resonance

...had a significant impact on garnering and sustaining user attention. Optimizing these creative dimensions represented an opportunity to amplify brand impact even further. The findings from this large-scale meta-analysis provide robust evidence that attention metrics like active viewable time are far more predictive of actual brand outcomes compared to standard digital ad delivery metrics.

As the researchers summarize, "The results show a clear and direct correlation between attention and brand effects...a viewable impression alone is not enough to drive brand lift. Attention is a key ingredient.

"For brands, this study highlights the importance of prioritizing ad placements, formats, and creative strategies that foster active consumer attention and engagement. Simply maximizing ad impressions and viewability is unlikely to move the needle on critical brand KPIs. As digital advertising continues to evolve with new ad tech and formats, metrics that quantify actual human attention will likely become even more crucial for maximizing advertising effectiveness and ROI.


Harbine Engineering Journal Study on Ad Impact


The study published in the Harbine Engineering Journal aimed to investigate the impact of advertisements on key metrics like brand awareness, customer satisfaction, and purchase intent. The researchers employed a multi-pronged approach that combined survey data with website analytics.


1. Survey Data Collection


The core of the study involved conducting surveys across three distinct groups:

  • Customers: A random sample of 5,000 customers who had made a purchase from the company's e-commerce website in the past 6 months.

  • Employees: 250 employees across various roles and departments within the company.

  • Management: 35 members of the senior leadership team and executive management.


The surveys captured respondents' perceptions of the company's advertising campaigns, brand messaging, and overall customer experience. Specific questions focused on:


  • Recalling and recognizing the company's ads across channels (TV, digital, print, etc.)

  • Perceived impact of ads on brand awareness and purchase decisions

  • Satisfaction with the website experience and e-commerce journey

  • Likelihood to recommend the brand (Net Promoter Score)


2. Website Analytics Integration


To complement the survey insights, the researchers also analyzed website behavior data captured via analytics tools like Google Analytics and Hotjar. Key metrics examined included:


  • Traffic sources (organic, paid, referral, etc.)

  • User journeys and conversion funnel performance

  • Engagement signals (pages visited, time on site, clicks, etc.)

  • Audience demographics (age, location, interests, etc.)


The website data allowed the researchers to quantify actual user behavior and correlate it with the stated perceptions from the surveys.


3. Statistical Analysis


The final stage involved rigorous statistical analysis to uncover potential relationships and drivers of advertising effectiveness. This included:

  • Correlation Analysis: Examining the correlation between survey responses (e.g. ad recall) and website behavior metrics to identify potential links.

  • Regression Modeling: Developing regression models to quantify the impact of factors like ad exposure, demographics, and website experience on key outcomes like satisfaction and purchase intent.

  • Multivariate Testing: Analyzing differences in survey responses and website behavior across variations in the advertising creative/messaging using techniques like A/B testing.


A key focus area was understanding if demographic factors like age had any bearing on the observed advertising impact and overall user experience.


Key Findings of the Harbine Engineering Journal Study on Ad Impact


The study yielded several noteworthy findings regarding the effectiveness of the company's advertising efforts:


1. No Correlation Between Age and Website Satisfaction


Contrary to common assumptions, the analysis revealed no statistically significant correlation between a user's age and their stated satisfaction with the website experience. Both younger and older demographics exhibited similar levels of satisfaction, suggesting the website design and user experience were well-tailored to a broad audience.


2. Opportunities to Improve Ad Visibility and Targeting


While overall ad recall was relatively high (68% for customers), the surveys highlighted opportunities to increase visibility and resonance of the advertising creative. Employees and management tended to overestimate the impact and memorability of the company's ads compared to actual customer perceptions. This misalignment signaled a need for more rigorous market testing and persona-based targeting of the advertising campaigns.


3. Positive Correlation Between Ad Exposure and Purchase Intent


One of the most encouraging findings was a statistically significant positive correlation between survey respondents' recollection of seeing the company's ads and their stated likelihood to make a purchase. Customers who reported higher ad exposure and recall also tended to exhibit higher purchase intent scores and were more likely to recommend the brand.


4. Importance of Seamless User Experience


The study found that a seamless, intuitive user experience on the company's website and e-commerce platform was crucial for maximizing the impact of advertising efforts. Respondents who reported frustrations or pain points during the purchase journey were far less likely to convert, regardless of their initial ad exposure and awareness levels. This underscored the need for consistent quality across all touchpoints in the marketing and sales funnel. While the study confirmed the overall positive impact of the company's advertising campaigns, it also highlighted several areas for optimization and improvement:


  • Enhancing ad creative and messaging to boost visibility and memorability

  • Leveraging more advanced audience segmentation and targeting capabilities

  • Maintaining a relentless focus on delivering seamless user experiences

  • Closing perception gaps between internal stakeholders and customers


By combining robust survey data with actual user behavior insights, this study provided a comprehensive view into the real-world effectiveness of advertising initiatives. The findings can inform more impactful, ROI-driven strategies for allocating ad budgets and orchestrating cohesive cross-channel experiences.


Columbia Business School: Factors influencing when advertising impacts brand awareness versus purchase intent


The Columbia Business School meta-analysis aimed to synthesize research findings on the key factors that influence when advertising impacts brand awareness versus purchase intent. The researchers conducted a comprehensive review of over 200 papers published in major marketing journals between 1990-2012.

Studies were included in the meta-analysis if they met the following criteria:


  1. Empirically measured the impact of advertising on either brand awareness metrics (e.g. recall, recognition) or purchase-related metrics (e.g. intent, choice, sales)

  2. Reported statistical estimates of advertising's effect size

  3. Provided details on key moderating variables like product category, ad content factors, media factors, etc.


This rigorous screening process yielded a final sample of 57 studies spanning various product categories, countries, time periods, and methodologies (experiments, surveys, econometric models, etc.). The authors then coded each study along multiple dimensions:


  • Dependent variable type (awareness vs. purchase)

  • Product category (CPG, durables, services, etc.)

  • Advertising content factors (rational vs emotional appeals, comparative ads, etc.)

  • Advertising execution factors (repetition, media vehicle, spend/weight, etc.)

  • Brand characteristics (quality, differentiation, market share, etc.)

  • Methodological characteristics (data source, estimation technique, etc.)


Meta-Analytic Approach


The analysis involved calculating separate meta-analytic means of the advertising effect sizes for:


  1. Brand awareness metrics

  2. Purchase-related metrics


This allowed comparing the overall magnitude of advertising's impact on each outcome type. The researchers also used meta-regression to examine how the different coded variables moderated advertising's effect on awareness versus purchase metrics. Weighted least squares regressions were specified with:


  • Effect size as the dependent variable

  • Coded study characteristics as predictors

  • Weights based on the inverse of each effect's variance


This approach quantified the influence of factors like product category, ad appeals, media factors, brand equity, etc. on advertising effectiveness.


Key Findings of the Columbia Business School Meta Analysis on Factors Influencing when Advertising Impacts Brand Awareness versus Purchase Intent


The meta-analysis yielded several important insights into the differential effects of advertising on building awareness versus driving purchase behavior:


1. Larger Effects on Awareness than Purchase Intent


Across the studies analyzed, advertising had a significantly larger overall effect on enhancing brand awareness metrics compared to impacting purchase-related outcomes like choice and sales. The mean effect size for awareness was 0.32, versus just 0.16 for purchase metrics - about half as large.


2. Moderating Role of Advertising Weight


As expected, higher advertising weight (measured by spend levels or repetition) increased advertising's ability to build awareness. However, it had a smaller positive impact on purchase metrics. This suggests that heavy ad spending is more effective for achieving widespread brand salience versus directly driving purchase behavior.


3. Importance of Brand Equity


The analysis found that a brand's existing equity level was a key moderator of advertising effectiveness:


  • For high-equity brands, advertising had a stronger positive effect on purchase metrics

  • For low-equity brands, advertising provided a bigger boost to building awareness


This highlights how brands may need to first invest in awareness-building before ads can effectively influence purchase.


4. Content Effects on Awareness vs. Purchase


The use of emotional ad appeals and comparative advertising claims enhanced advertising's impact on purchase metrics more than awareness. Conversely, repetition of similar executions was more effective for reinforcing brand awareness versus driving purchase intentions.


5. Media Vehicle Differences


The meta-regression revealed that different media vehicles had varying effects:


  • Television had the largest positive impact on awareness

  • Print media like magazines were most effective for influencing purchase

  • Digital media had smaller effects than expected on both outcomes


However, the authors note this may partly reflect the time period studied (pre-2012) when digital ad spend and measurement were much more limited.


Overall, this comprehensive meta-analysis provides a nuanced perspective on how various factors moderate whether advertising achieves brand-building objectives versus more direct sales impacts.


The findings can help guide strategic decisions around creative content, media planning, and budgeting based on specific brand/marketing objectives. They also highlight the importance of accounting for a brand's existing equity when setting advertising expectations and measuring effectiveness.


Key Insights from the Studies Summarized


Harvard Study on Search and Display Ad Interaction:

  • Display ads significantly increase paid search conversions, but with a 2-week delay

  • Both search and display ads exhibit strong positive carryover/dynamic effects over time

  • Display ads also increase search ad clicks and costs, offsetting some of the conversion lift

  • Accounting for attribution to search, dynamics, and added costs alters performance metrics

  • The optimal budget allocation should increase search's share by up to 36% despite display's attribution benefit

Nielsen Study on New Media's Brand Impact:

  • Median brand awareness lift of 70% from podcast/influencer/branded content exposure

  • Meaningful lifts also seen for favorability (28%), purchase intent (18%), short-term ROI (2.7x)

  • High variance - top quartile saw 150%+ awareness lift, bottom quartile was negligible

  • Attention and viewability were key drivers of effectiveness


Lumen/Teads/Dynata Study on Attention and Brand Effects:

  • Clear correlation between attentive ad exposure and upper-funnel brand metrics

  • Lower-funnel metrics like purchase intent saw steep lift between 5-10 attentive seconds

  • Attentive impressions drove 2-4x higher brand lift than just viewable impressions

  • Creative optimization can further amplify attention and brand impact


Harbine Engineering Journal Study on Ad Impact:

  • No correlation between age and website satisfaction, suggesting good user experience

  • Opportunities to improve ad visibility, targeting, and alignment with customer perceptions

  • Positive correlation between ad exposure and purchase intent

  • Seamless user experience crucial for maximizing ad effectiveness


Columbia Business School Meta-Analysis:

  • Advertising had larger effect on building awareness vs. driving purchase metrics

  • Higher ad weight boosted awareness more for high-equity brands

  • Emotional appeals and comparative ads enhanced purchase impact

  • TV was best for awareness, print was best for purchase influence


The studies highlight the importance of accounting for cross-channel attribution effects and carryover dynamics when evaluating online advertising performance.

Display ads were found to significantly increase search ad conversions, but with a delayed impact, while both channels exhibited strong positive dynamics that improved effectiveness over time.

However, display ads also drove up search costs. Accounting for these attribution and dynamic effects substantially altered performance metrics like revenue per ad dollar spent.


The optimal budget allocation favored increasing the search budget share despite display's attribution benefit, due to search's stronger dynamics. Brands should prioritize ad placements and creative that foster active consumer attention and engagement to maximize impact on critical upper-funnel metrics like awareness and favorability.


Leveraging advanced measurement techniques, predictive analytics, and an integrated cross-channel strategy can help marketers navigate attribution challenges, optimize media budgets, and orchestrate cohesive experiences that drive awareness and conversions.



Sources

  1. Kireyev, P., Pauwels, K., & Gupta, S. (2013). Do display ads influence search? Attribution and dynamics in online advertising. Harvard Business School Working Paper, No. 13-070.

  2. Nielsen (2022). Measuring new media's impact on brand awareness and ROI. https://www.nielsen.com/insights/2022/measuring-new-medias-impact-on-brand-awareness-and-roi/

  3. WARC (2024). New study shows a clear link between attention and brand effects. https://www.warc.com/content/feed/new-study-shows-a-clear-link-between-attention-and-brand-effects/en-GB/8527

  4. Harbine Engineering Journal (2024). An investigation into the impact of ads on brand awareness. https://harbinengineeringjournal.com/index.php/journal/article/view/2280

  5. Columbia Business School (2013). Factors influencing when advertising impacts brand awareness versus purchase intent: A meta-analysis.

  6. Faster Capital (2024). Exploring the latest research on enhancing brand awareness strategies. https://fastercapital.com/content/Exploring-latest-research-on-brand-awareness-strategies.html

Commentaires